xen_internal.c 50.6 KB
Newer Older
1 2 3
/*
 * xen_internal.c: direct access to Xen hypervisor level
 *
D
Daniel Veillard 已提交
4
 * Copyright (C) 2005, 2006 Red Hat, Inc.
5 6 7 8 9 10 11 12
 *
 * See COPYING.LIB for the License of this software
 *
 * Daniel Veillard <veillard@redhat.com>
 */

#include <stdio.h>
#include <string.h>
13
/* required for uint8_t, uint32_t, etc ... */
14 15 16 17 18 19 20
#include <stdint.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/mman.h>
#include <sys/ioctl.h>
21
#include <limits.h>
22 23 24
#include <stdint.h>

/* required for dom0_getdomaininfo_t */
25
#include <xen/dom0_ops.h>
26
#include <xen/version.h>
27
#include <xen/xen.h>
28
#include <xen/linux/privcmd.h>
29

30 31 32
/* required for shutdown flags */
#include <xen/sched.h>

33 34
/* #define DEBUG */
/*
35
 * so far there is 2 versions of the structures usable for doing
36 37 38 39
 * hypervisor calls.
 */
/* the old one */
typedef struct v0_hypercall_struct {
40 41
    unsigned long op;
    unsigned long arg[5];
42 43 44 45 46 47 48 49 50 51
} v0_hypercall_t;
#define XEN_V0_IOCTL_HYPERCALL_CMD \
        _IOC(_IOC_NONE, 'P', 0, sizeof(v0_hypercall_t))

/* the new one */
typedef struct v1_hypercall_struct
{
    uint64_t op;
    uint64_t arg[5];
} v1_hypercall_t;
52 53
#define XEN_V1_IOCTL_HYPERCALL_CMD                  \
    _IOC(_IOC_NONE, 'P', 0, sizeof(v1_hypercall_t))
54

55 56 57 58 59 60 61 62
typedef v1_hypercall_t hypercall_t;

#ifndef __HYPERVISOR_sysctl
#define __HYPERVISOR_sysctl 35
#endif
#ifndef __HYPERVISOR_domctl
#define __HYPERVISOR_domctl 36
#endif
63

64 65
static int xen_ioctl_hypercall_cmd = 0;
static int initialized = 0;
66
static int in_init = 0;
67
static int hv_version = 0;
68 69 70
static int hypervisor_version = 2;
static int sys_interface_version = -1;
static int dom_interface_version = -1;
71
static int kb_per_pages = 0;
72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87

/*
 * The content of the structures for a getdomaininfolist system hypercall
 */
#ifndef DOMFLAGS_DYING
#define DOMFLAGS_DYING     (1<<0) /* Domain is scheduled to die.             */
#define DOMFLAGS_SHUTDOWN  (1<<2) /* The guest OS has shut down.             */
#define DOMFLAGS_PAUSED    (1<<3) /* Currently paused by control software.   */
#define DOMFLAGS_BLOCKED   (1<<4) /* Currently blocked pending an event.     */
#define DOMFLAGS_RUNNING   (1<<5) /* Domain is currently running.            */
#define DOMFLAGS_CPUMASK      255 /* CPU to which this domain is bound.      */
#define DOMFLAGS_CPUSHIFT       8
#define DOMFLAGS_SHUTDOWNMASK 255 /* DOMFLAGS_SHUTDOWN guest-supplied code.  */
#define DOMFLAGS_SHUTDOWNSHIFT 16
#endif

88 89 90 91 92 93 94 95 96 97 98
/*
 * These flags explain why a system is in the state of "shutdown".  Normally,
 * They are defined in xen/sched.h
 */
#ifndef SHUTDOWN_poweroff
#define SHUTDOWN_poweroff   0  /* Domain exited normally. Clean up and kill. */
#define SHUTDOWN_reboot     1  /* Clean up, kill, and then restart.          */
#define SHUTDOWN_suspend    2  /* Clean up, save suspend info, kill.         */
#define SHUTDOWN_crash      3  /* Tell controller we've crashed.             */
#endif

99 100 101 102 103 104 105 106 107
#define XEN_V0_OP_GETDOMAININFOLIST	38
#define XEN_V1_OP_GETDOMAININFOLIST	38
#define XEN_V2_OP_GETDOMAININFOLIST	6

struct xen_v0_getdomaininfo {
    domid_t  domain;	/* the domain number */
    uint32_t flags;	/* falgs, see before */
    uint64_t tot_pages;	/* total number of pages used */
    uint64_t max_pages;	/* maximum number of pages allowed */
108
    unsigned long shared_info_frame; /* MFN of shared_info struct */
109 110 111 112 113 114 115 116
    uint64_t cpu_time;  /* CPU time used */
    uint32_t nr_online_vcpus;  /* Number of VCPUs currently online. */
    uint32_t max_vcpu_id; /* Maximum VCPUID in use by this domain. */
    uint32_t ssidref;
    xen_domain_handle_t handle;
};
typedef struct xen_v0_getdomaininfo xen_v0_getdomaininfo;

117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162
struct xen_v2_getdomaininfo {
    domid_t  domain;	/* the domain number */
    uint32_t flags;	/* falgs, see before */
    uint64_t tot_pages;	/* total number of pages used */
    uint64_t max_pages;	/* maximum number of pages allowed */
    uint64_t shared_info_frame; /* MFN of shared_info struct */
    uint64_t cpu_time;  /* CPU time used */
    uint32_t nr_online_vcpus;  /* Number of VCPUs currently online. */
    uint32_t max_vcpu_id; /* Maximum VCPUID in use by this domain. */
    uint32_t ssidref;
    xen_domain_handle_t handle;
};
typedef struct xen_v2_getdomaininfo xen_v2_getdomaininfo;

union xen_getdomaininfo {
    struct xen_v0_getdomaininfo v0;
    struct xen_v2_getdomaininfo v2;
};
typedef union xen_getdomaininfo xen_getdomaininfo;

union xen_getdomaininfolist {
    struct xen_v0_getdomaininfo *v0;
    struct xen_v2_getdomaininfo *v2;
};
typedef union xen_getdomaininfolist xen_getdomaininfolist;

#define XEN_GETDOMAININFOLIST_ALLOC(domlist, size)                      \
    (hypervisor_version < 2 ?                                           \
     ((domlist.v0 = malloc(sizeof(xen_v0_getdomaininfo)*(size))) != NULL) : \
     ((domlist.v2 = malloc(sizeof(xen_v2_getdomaininfo)*(size))) != NULL))

#define XEN_GETDOMAININFOLIST_FREE(domlist)     \
    (hypervisor_version < 2 ?                   \
     free(domlist.v0) :                         \
     free(domlist.v2))

#define XEN_GETDOMAININFOLIST_CLEAR(domlist, size)                  \
    (hypervisor_version < 2 ?                                       \
     memset(domlist.v0, 0, sizeof(xen_v0_getdomaininfo) * size) :   \
     memset(domlist.v2, 0, sizeof(xen_v2_getdomaininfo) * size))

#define XEN_GETDOMAININFOLIST_DOMAIN(domlist, n)    \
    (hypervisor_version < 2 ?                       \
     domlist.v0[n].domain :                         \
     domlist.v2[n].domain)

163 164 165 166
#define XEN_GETDOMAININFOLIST_DATA(domlist)     \
    (hypervisor_version < 2 ?                   \
     (void*)(domlist->v0) :                     \
     (void*)(domlist->v2))
167

168 169 170 171
#define XEN_GETDOMAININFO_SIZE                  \
    (hypervisor_version < 2 ?                   \
     sizeof(xen_v0_getdomaininfo) :             \
     sizeof(xen_v2_getdomaininfo))
172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210

#define XEN_GETDOMAININFO_CLEAR(dominfo)                        \
    (hypervisor_version < 2 ?                                   \
     memset(&(dominfo.v0), 0, sizeof(xen_v0_getdomaininfo)) :   \
     memset(&(dominfo.v2), 0, sizeof(xen_v2_getdomaininfo)))

#define XEN_GETDOMAININFO_DOMAIN(dominfo)       \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.domain :                        \
     dominfo.v2.domain)

#define XEN_GETDOMAININFO_CPUTIME(dominfo)      \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.cpu_time :                      \
     dominfo.v2.cpu_time)

#define XEN_GETDOMAININFO_CPUCOUNT(dominfo)     \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.nr_online_vcpus :               \
     dominfo.v2.nr_online_vcpus)

#define XEN_GETDOMAININFO_FLAGS(dominfo)        \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.flags :                         \
     dominfo.v2.flags)

#define XEN_GETDOMAININFO_TOT_PAGES(dominfo)    \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.tot_pages :                     \
     dominfo.v2.tot_pages)

#define XEN_GETDOMAININFO_MAX_PAGES(dominfo)    \
    (hypervisor_version < 2 ?                   \
     dominfo.v0.max_pages :                     \
     dominfo.v2.max_pages)



struct xen_v0_getdomaininfolistop {
211 212 213 214 215
    domid_t   first_domain;
    uint32_t  max_domains;
    struct xen_v0_getdomaininfo *buffer;
    uint32_t  num_domains;
};
216 217 218 219 220 221 222 223 224 225 226 227
typedef struct xen_v0_getdomaininfolistop xen_v0_getdomaininfolistop;


struct xen_v2_getdomaininfolistop {
    domid_t   first_domain;
    uint32_t  max_domains;
    struct xen_v2_getdomaininfo *buffer;
    uint32_t  num_domains;
};
typedef struct xen_v2_getdomaininfolistop xen_v2_getdomaininfolistop;


228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319

struct xen_v0_domainop {
    domid_t   domain;
};
typedef struct xen_v0_domainop xen_v0_domainop;

/*
 * The informations for a destroydomain system hypercall
 */
#define XEN_V0_OP_DESTROYDOMAIN	9
#define XEN_V1_OP_DESTROYDOMAIN	9
#define XEN_V2_OP_DESTROYDOMAIN	2

/*
 * The informations for a pausedomain system hypercall
 */
#define XEN_V0_OP_PAUSEDOMAIN	10
#define XEN_V1_OP_PAUSEDOMAIN	10
#define XEN_V2_OP_PAUSEDOMAIN	3

/*
 * The informations for an unpausedomain system hypercall
 */
#define XEN_V0_OP_UNPAUSEDOMAIN	11
#define XEN_V1_OP_UNPAUSEDOMAIN	11
#define XEN_V2_OP_UNPAUSEDOMAIN	4

/*
 * The informations for an setmaxmem system hypercall
 */
#define XEN_V0_OP_SETMAXMEM	28
#define XEN_V1_OP_SETMAXMEM	28
#define XEN_V2_OP_SETMAXMEM	14

struct xen_v0_setmaxmem {
    domid_t	domain;
    uint64_t	maxmem;
};
typedef struct xen_v0_setmaxmem xen_v0_setmaxmem;
typedef struct xen_v0_setmaxmem xen_v1_setmaxmem;

struct xen_v2_setmaxmem {
    uint64_t	maxmem;
};
typedef struct xen_v2_setmaxmem xen_v2_setmaxmem;

/*
 * The informations for an setmaxvcpu system hypercall
 */
#define XEN_V0_OP_SETMAXVCPU	41
#define XEN_V1_OP_SETMAXVCPU	41
#define XEN_V2_OP_SETMAXVCPU	15

struct xen_v0_setmaxvcpu {
    domid_t	domain;
    uint32_t	maxvcpu;
};
typedef struct xen_v0_setmaxvcpu xen_v0_setmaxvcpu;
typedef struct xen_v0_setmaxvcpu xen_v1_setmaxvcpu;

struct xen_v2_setmaxvcpu {
    uint32_t	maxvcpu;
};
typedef struct xen_v2_setmaxvcpu xen_v2_setmaxvcpu;

/*
 * The informations for an setvcpumap system hypercall
 * Note that between 1 and 2 the limitation to 64 physical CPU was lifted
 * hence the difference in structures
 */
#define XEN_V0_OP_SETVCPUMAP	20
#define XEN_V1_OP_SETVCPUMAP	20
#define XEN_V2_OP_SETVCPUMAP	9

struct xen_v0_setvcpumap {
    domid_t	domain;
    uint32_t	vcpu;
    cpumap_t    cpumap;
};
typedef struct xen_v0_setvcpumap xen_v0_setvcpumap;
typedef struct xen_v0_setvcpumap xen_v1_setvcpumap;

struct xen_v2_cpumap {
    uint8_t    *bitmap;
    uint32_t    nr_cpus;
};
struct xen_v2_setvcpumap {
    uint32_t	vcpu;
    struct xen_v2_cpumap cpumap;
};
typedef struct xen_v2_setvcpumap xen_v2_setvcpumap;

320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356
/*
 * The informations for an vcpuinfo system hypercall
 */
#define XEN_V0_OP_GETVCPUINFO   43
#define XEN_V1_OP_GETVCPUINFO	43
#define XEN_V2_OP_GETVCPUINFO   14

struct xen_v0_vcpuinfo {
    domid_t	domain;		/* owner's domain */
    uint32_t	vcpu;		/* the vcpu number */
    uint8_t	online;		/* seen as on line */
    uint8_t	blocked;	/* blocked on event */
    uint8_t	running;	/* scheduled on CPU */
    uint64_t    cpu_time;	/* nanosecond of CPU used */
    uint32_t	cpu;		/* current mapping */
    cpumap_t	cpumap;		/* deprecated in V2 */
};
typedef struct xen_v0_vcpuinfo xen_v0_vcpuinfo;
typedef struct xen_v0_vcpuinfo xen_v1_vcpuinfo;

struct xen_v2_vcpuinfo {
    domid_t	domain;		/* owner's domain */
    uint32_t	vcpu;		/* the vcpu number */
    uint8_t	online;		/* seen as on line */
    uint8_t	blocked;	/* blocked on event */
    uint8_t	running;	/* scheduled on CPU */
    uint64_t    cpu_time;	/* nanosecond of CPU used */
    uint32_t	cpu;		/* current mapping */
};
typedef struct xen_v2_vcpuinfo xen_v2_vcpuinfo;

/*
 * from V2 the pinning of a vcpu is read with a separate call
 */
#define XEN_V2_OP_GETVCPUMAP	25
typedef struct xen_v2_setvcpumap xen_v2_getvcpumap;

357 358 359 360 361 362 363 364 365
/*
 * The hypercall operation structures also have changed on
 * changeset 86d26e6ec89b
 */
/* the old structure */
struct xen_op_v0 {
    uint32_t cmd;
    uint32_t interface_version;
    union {
366 367 368 369 370 371 372
        xen_v0_getdomaininfolistop getdomaininfolist;
        xen_v0_domainop          domain;
        xen_v0_setmaxmem         setmaxmem;
        xen_v0_setmaxvcpu        setmaxvcpu;
        xen_v0_setvcpumap        setvcpumap;
        xen_v0_vcpuinfo          getvcpuinfo;
        uint8_t padding[128];
373 374 375 376 377 378 379 380 381 382
    } u;
};
typedef struct xen_op_v0 xen_op_v0;
typedef struct xen_op_v0 xen_op_v1;

/* the new structure for systems operations */
struct xen_op_v2_sys {
    uint32_t cmd;
    uint32_t interface_version;
    union {
383 384
        xen_v2_getdomaininfolistop getdomaininfolist;
        uint8_t padding[128];
385 386 387 388 389 390 391 392 393 394
    } u;
};
typedef struct xen_op_v2_sys xen_op_v2_sys;

/* the new structure for domains operation */
struct xen_op_v2_dom {
    uint32_t cmd;
    uint32_t interface_version;
    domid_t  domain;
    union {
395 396 397 398 399 400
        xen_v2_setmaxmem         setmaxmem;
        xen_v2_setmaxvcpu        setmaxvcpu;
        xen_v2_setvcpumap        setvcpumap;
        xen_v2_vcpuinfo          getvcpuinfo;
        xen_v2_getvcpumap        getvcpumap;
        uint8_t padding[128];
401 402 403
    } u;
};
typedef struct xen_op_v2_dom xen_op_v2_dom;
404 405

#include "internal.h"
406
#include "driver.h"
407 408 409 410
#include "xen_internal.h"

#define XEN_HYPERVISOR_SOCKET "/proc/xen/privcmd"

411
#ifndef PROXY
412
static const char * xenHypervisorGetType(virConnectPtr conn);
413
static unsigned long xenHypervisorGetMaxMemory(virDomainPtr domain);
414
#endif
415
static int xenHypervisorInit(void);
416

417
#ifndef PROXY
418
static virDriver xenHypervisorDriver = {
419
    VIR_DRV_XEN_HYPERVISOR,
420
    "Xen",
421 422 423
    (DOM0_INTERFACE_VERSION >> 24) * 1000000 +
    ((DOM0_INTERFACE_VERSION >> 16) & 0xFF) * 1000 +
    (DOM0_INTERFACE_VERSION & 0xFFFF),
424
    xenHypervisorInit, /* init */
425 426
    xenHypervisorOpen, /* open */
    xenHypervisorClose, /* close */
427
    xenHypervisorGetType, /* type */
428
    xenHypervisorGetVersion, /* version */
429
    NULL, /* nodeGetInfo */
430 431
    xenHypervisorListDomains, /* listDomains */
    xenHypervisorNumOfDomains, /* numOfDomains */
432 433 434 435 436 437 438
    NULL, /* domainCreateLinux */
    NULL, /* domainLookupByID */
    NULL, /* domainLookupByUUID */
    NULL, /* domainLookupByName */
    xenHypervisorPauseDomain, /* domainSuspend */
    xenHypervisorResumeDomain, /* domainResume */
    NULL, /* domainShutdown */
439
    NULL, /* domainReboot */
440 441 442 443 444 445
    xenHypervisorDestroyDomain, /* domainDestroy */
    NULL, /* domainFree */
    NULL, /* domainGetName */
    NULL, /* domainGetID */
    NULL, /* domainGetUUID */
    NULL, /* domainGetOSType */
446
    xenHypervisorGetMaxMemory, /* domainGetMaxMemory */
447
    xenHypervisorSetMaxMemory, /* domainSetMaxMemory */
448
    NULL, /* domainSetMemory */
449 450
    xenHypervisorGetDomainInfo, /* domainGetInfo */
    NULL, /* domainSave */
451 452 453
    NULL, /* domainRestore */
    xenHypervisorSetVcpus, /* domainSetVcpus */
    xenHypervisorPinVcpu, /* domainPinVcpu */
454 455
    xenHypervisorGetVcpus, /* domainGetVcpus */
    NULL, /* domainDumpXML */
456 457 458 459 460
    NULL, /* listDefinedDomains */
    NULL, /* numOfDefinedDomains */
    NULL, /* domainCreate */
    NULL, /* domainDefineXML */
    NULL, /* domainUndefine */
461
};
462
#endif /* !PROXY */
463

464 465 466 467 468 469 470 471 472
/**
 * virXenError:
 * @conn: the connection if available
 * @error: the error number
 * @info: extra information string
 *
 * Handle an error at the xend daemon interface
 */
static void
473 474
virXenError(virErrorNumber error, const char *info, int value)
{
475
    const char *errmsg;
476

477
    if ((error == VIR_ERR_OK) || (in_init != 0))
478 479 480 481
        return;

    errmsg = __virErrorMsg(error, info);
    __virRaiseError(NULL, NULL, VIR_FROM_XEN, error, VIR_ERR_ERROR,
482
                    errmsg, info, NULL, value, 0, errmsg, info, value);
483 484
}

485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609
/**
 * xenHypervisorDoV0Op:
 * @handle: the handle to the Xen hypervisor
 * @op: pointer to the hyperviros operation structure
 *
 * Do an hypervisor operation though the old interface,
 * this leads to an hypervisor call through ioctl.
 *
 * Returns 0 in case of success and -1 in case of error.
 */
static int
xenHypervisorDoV0Op(int handle, xen_op_v0 * op)
{
    int ret;
    v0_hypercall_t hc;

    memset(&hc, 0, sizeof(hc));
    op->interface_version = hv_version << 8;
    hc.op = __HYPERVISOR_dom0_op;
    hc.arg[0] = (unsigned long) op;

    if (mlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " locking", sizeof(*op));
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
        virXenError(VIR_ERR_XEN_CALL, " ioctl ", xen_ioctl_hypercall_cmd);
    }

    if (munlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
        ret = -1;
    }

    if (ret < 0)
        return (-1);

    return (0);
}
/**
 * xenHypervisorDoV1Op:
 * @handle: the handle to the Xen hypervisor
 * @op: pointer to the hyperviros operation structure
 *
 * Do an hypervisor v1 operation, this leads to an hypervisor call through
 * ioctl.
 *
 * Returns 0 in case of success and -1 in case of error.
 */
static int
xenHypervisorDoV1Op(int handle, xen_op_v1* op)
{
    int ret;
    hypercall_t hc;

    memset(&hc, 0, sizeof(hc));
    op->interface_version = DOM0_INTERFACE_VERSION;
    hc.op = __HYPERVISOR_dom0_op;
    hc.arg[0] = (unsigned long) op;

    if (mlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " locking", sizeof(*op));
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
        virXenError(VIR_ERR_XEN_CALL, " ioctl ", xen_ioctl_hypercall_cmd);
    }

    if (munlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
        ret = -1;
    }

    if (ret < 0)
        return (-1);

    return (0);
}

/**
 * xenHypervisorDoV2Sys:
 * @handle: the handle to the Xen hypervisor
 * @op: pointer to the hypervisor operation structure
 *
 * Do an hypervisor v2 stsyem operation, this leads to an hypervisor
 * call through ioctl.
 *
 * Returns 0 in case of success and -1 in case of error.
 */
static int
xenHypervisorDoV2Sys(int handle, xen_op_v2_sys* op)
{
    int ret;
    hypercall_t hc;

    memset(&hc, 0, sizeof(hc));
    op->interface_version = sys_interface_version;
    hc.op = __HYPERVISOR_sysctl;
    hc.arg[0] = (unsigned long) op;

    if (mlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " locking", sizeof(*op));
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
        virXenError(VIR_ERR_XEN_CALL, " ioctl ", xen_ioctl_hypercall_cmd);
    }

    if (munlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
        ret = -1;
    }

    if (ret < 0)
        return (-1);

    return (0);
}

610
#ifndef PROXY
611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651
/**
 * xenHypervisorDoV2Dom:
 * @handle: the handle to the Xen hypervisor
 * @op: pointer to the hypervisor domain operation structure
 *
 * Do an hypervisor v2 domain operation, this leads to an hypervisor
 * call through ioctl.
 *
 * Returns 0 in case of success and -1 in case of error.
 */
static int
xenHypervisorDoV2Dom(int handle, xen_op_v2_dom* op)
{
    int ret;
    hypercall_t hc;

    memset(&hc, 0, sizeof(hc));
    op->interface_version = dom_interface_version;
    hc.op = __HYPERVISOR_domctl;
    hc.arg[0] = (unsigned long) op;

    if (mlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " locking", sizeof(*op));
        return (-1);
    }

    ret = ioctl(handle, xen_ioctl_hypercall_cmd, (unsigned long) &hc);
    if (ret < 0) {
        virXenError(VIR_ERR_XEN_CALL, " ioctl ", xen_ioctl_hypercall_cmd);
    }

    if (munlock(op, sizeof(dom0_op_t)) < 0) {
        virXenError(VIR_ERR_XEN_CALL, " releasing", sizeof(*op));
        ret = -1;
    }

    if (ret < 0)
        return (-1);

    return (0);
}
652
#endif /* PROXY */
653 654 655 656 657 658 659 660 661 662 663 664 665 666

/**
 * virXen_getdomaininfolist:
 * @handle: the hypervisor handle
 * @first_domain: first domain in the range
 * @maxids: maximum number of domains to list
 * @dominfos: output structures
 *
 * Do a low level hypercall to list existing domains informations
 *
 * Returns the number of domains or -1 in case of failure
 */
static int
virXen_getdomaininfolist(int handle, int first_domain, int maxids,
667
                         xen_getdomaininfolist *dominfos)
668 669 670
{
    int ret = -1;

671 672
    if (mlock(XEN_GETDOMAININFOLIST_DATA(dominfos),
              XEN_GETDOMAININFO_SIZE * maxids) < 0) {
673
        virXenError(VIR_ERR_XEN_CALL, " locking",
674
                    XEN_GETDOMAININFO_SIZE * maxids);
675 676 677 678 679 680
        return (-1);
    }
    if (hypervisor_version > 1) {
        xen_op_v2_sys op;

        memset(&op, 0, sizeof(op));
681 682 683 684 685 686 687 688
        op.cmd = XEN_V2_OP_GETDOMAININFOLIST;
        op.u.getdomaininfolist.first_domain = (domid_t) first_domain;
        op.u.getdomaininfolist.max_domains = maxids;
        op.u.getdomaininfolist.buffer = dominfos->v2;
        op.u.getdomaininfolist.num_domains = maxids;
        ret = xenHypervisorDoV2Sys(handle, &op);
        if (ret == 0)
            ret = op.u.getdomaininfolist.num_domains;
689 690 691 692
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
693 694 695 696 697 698 699 700
        op.cmd = XEN_V1_OP_GETDOMAININFOLIST;
        op.u.getdomaininfolist.first_domain = (domid_t) first_domain;
        op.u.getdomaininfolist.max_domains = maxids;
        op.u.getdomaininfolist.buffer = dominfos->v0;
        op.u.getdomaininfolist.num_domains = maxids;
        ret = xenHypervisorDoV1Op(handle, &op);
        if (ret == 0)
            ret = op.u.getdomaininfolist.num_domains;
701 702 703 704
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
705 706 707 708 709 710 711 712
        op.cmd = XEN_V0_OP_GETDOMAININFOLIST;
        op.u.getdomaininfolist.first_domain = (domid_t) first_domain;
        op.u.getdomaininfolist.max_domains = maxids;
        op.u.getdomaininfolist.buffer = dominfos->v0;
        op.u.getdomaininfolist.num_domains = maxids;
        ret = xenHypervisorDoV0Op(handle, &op);
        if (ret == 0)
            ret = op.u.getdomaininfolist.num_domains;
713
    }
714 715
    if (munlock(XEN_GETDOMAININFOLIST_DATA(dominfos),
                XEN_GETDOMAININFO_SIZE * maxids) < 0) {
716
        virXenError(VIR_ERR_XEN_CALL, " release",
717
                    XEN_GETDOMAININFO_SIZE * maxids);
718 719 720 721 722
        ret = -1;
    }
    return(ret);
}

723 724 725 726 727 728 729 730 731 732 733 734 735 736 737
static int
virXen_getdomaininfo(int handle, int first_domain,
                     xen_getdomaininfo *dominfo) {
    xen_getdomaininfolist dominfos;

    if (hypervisor_version < 2) {
        dominfos.v0 = &(dominfo->v0);
    } else {
        dominfos.v2 = &(dominfo->v2);
    }

    return virXen_getdomaininfolist(handle, first_domain, 1, &dominfos);
}


738
#ifndef PROXY
739 740 741 742 743 744 745 746 747 748
/**
 * virXen_pausedomain:
 * @handle: the hypervisor handle
 * @id: the domain id
 *
 * Do a low level hypercall to pause the domain
 *
 * Returns 0 or -1 in case of failure
 */
static int
749
virXen_pausedomain(int handle, int id)
750 751 752 753 754 755 756
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
757 758 759
        op.cmd = XEN_V2_OP_PAUSEDOMAIN;
        op.domain = (domid_t) id;
        ret = xenHypervisorDoV2Dom(handle, &op);
760 761 762 763
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
764 765 766
        op.cmd = XEN_V1_OP_PAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV1Op(handle, &op);
767 768 769 770
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
771 772 773
        op.cmd = XEN_V0_OP_PAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV0Op(handle, &op);
774 775 776 777 778 779 780 781 782 783 784 785 786 787
    }
    return(ret);
}

/**
 * virXen_unpausedomain:
 * @handle: the hypervisor handle
 * @id: the domain id
 *
 * Do a low level hypercall to unpause the domain
 *
 * Returns 0 or -1 in case of failure
 */
static int
788
virXen_unpausedomain(int handle, int id)
789 790 791 792 793 794 795
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
796 797 798
        op.cmd = XEN_V2_OP_UNPAUSEDOMAIN;
        op.domain = (domid_t) id;
        ret = xenHypervisorDoV2Dom(handle, &op);
799 800 801 802
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
803 804 805
        op.cmd = XEN_V1_OP_UNPAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV1Op(handle, &op);
806 807 808 809
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
810 811 812
        op.cmd = XEN_V0_OP_UNPAUSEDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV0Op(handle, &op);
813 814 815 816 817 818 819 820 821 822 823 824 825 826
    }
    return(ret);
}

/**
 * virXen_destroydomain:
 * @handle: the hypervisor handle
 * @id: the domain id
 *
 * Do a low level hypercall to destroy the domain
 *
 * Returns 0 or -1 in case of failure
 */
static int
827
virXen_destroydomain(int handle, int id)
828 829 830 831 832 833 834
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
835 836 837
        op.cmd = XEN_V2_OP_DESTROYDOMAIN;
        op.domain = (domid_t) id;
        ret = xenHypervisorDoV2Dom(handle, &op);
838 839 840 841
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
842 843 844
        op.cmd = XEN_V1_OP_DESTROYDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV1Op(handle, &op);
845 846 847 848
    } else if (hypervisor_version == 0) {
        xen_op_v0 op;

        memset(&op, 0, sizeof(op));
849 850 851
        op.cmd = XEN_V0_OP_DESTROYDOMAIN;
        op.u.domain.domain = (domid_t) id;
        ret = xenHypervisorDoV0Op(handle, &op);
852 853 854 855 856 857 858 859 860 861 862 863 864 865 866
    }
    return(ret);
}

/**
 * virXen_setmaxmem:
 * @handle: the hypervisor handle
 * @id: the domain id
 * @memory: the amount of memory in kilobytes
 *
 * Do a low level hypercall to change the max memory amount
 *
 * Returns 0 or -1 in case of failure
 */
static int
867
virXen_setmaxmem(int handle, int id, unsigned long memory)
868 869 870 871 872 873 874
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
875 876 877 878
        op.cmd = XEN_V2_OP_SETMAXMEM;
        op.domain = (domid_t) id;
        op.u.setmaxmem.maxmem = memory;
        ret = xenHypervisorDoV2Dom(handle, &op);
879 880 881 882
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
883 884 885 886
        op.cmd = XEN_V1_OP_SETMAXMEM;
        op.u.setmaxmem.domain = (domid_t) id;
        op.u.setmaxmem.maxmem = memory;
        ret = xenHypervisorDoV1Op(handle, &op);
887
    } else if (hypervisor_version == 0) {
888
        xen_op_v0 op;
889 890

        memset(&op, 0, sizeof(op));
891 892 893 894
        op.cmd = XEN_V0_OP_SETMAXMEM;
        op.u.setmaxmem.domain = (domid_t) id;
        op.u.setmaxmem.maxmem = memory;
        ret = xenHypervisorDoV0Op(handle, &op);
895 896 897 898 899 900 901 902 903 904 905 906 907 908 909
    }
    return(ret);
}

/**
 * virXen_setmaxvcpus:
 * @handle: the hypervisor handle
 * @id: the domain id
 * @vcpus: the numbers of vcpus
 *
 * Do a low level hypercall to change the max vcpus amount
 *
 * Returns 0 or -1 in case of failure
 */
static int
910
virXen_setmaxvcpus(int handle, int id, unsigned int vcpus)
911 912 913 914 915 916 917
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
918 919 920 921
        op.cmd = XEN_V2_OP_SETMAXVCPU;
        op.domain = (domid_t) id;
        op.u.setmaxvcpu.maxvcpu = vcpus;
        ret = xenHypervisorDoV2Dom(handle, &op);
922 923 924 925
    } else if (hypervisor_version == 1) {
        xen_op_v1 op;

        memset(&op, 0, sizeof(op));
926 927 928 929
        op.cmd = XEN_V1_OP_SETMAXVCPU;
        op.u.setmaxvcpu.domain = (domid_t) id;
        op.u.setmaxvcpu.maxvcpu = vcpus;
        ret = xenHypervisorDoV1Op(handle, &op);
930
    } else if (hypervisor_version == 0) {
931
        xen_op_v0 op;
932 933

        memset(&op, 0, sizeof(op));
934 935 936 937
        op.cmd = XEN_V0_OP_SETMAXVCPU;
        op.u.setmaxvcpu.domain = (domid_t) id;
        op.u.setmaxvcpu.maxvcpu = vcpus;
        ret = xenHypervisorDoV0Op(handle, &op);
938 939 940 941 942 943 944 945 946 947
    }
    return(ret);
}

/**
 * virXen_setvcpumap:
 * @handle: the hypervisor handle
 * @id: the domain id
 * @vcpu: the vcpu to map
 * @cpumap: the bitmap for this vcpu
948
 * @maplen: the size of the bitmap in bytes
949 950 951 952 953 954 955 956 957 958 959 960 961 962
 *
 * Do a low level hypercall to change the pinning for vcpu
 *
 * Returns 0 or -1 in case of failure
 */
static int
virXen_setvcpumap(int handle, int id, unsigned int vcpu,
                  unsigned char * cpumap, int maplen)
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

963 964 965 966
        if (mlock(cpumap, maplen) < 0) {
            virXenError(VIR_ERR_XEN_CALL, " locking", maplen);
            return (-1);
        }
967
        memset(&op, 0, sizeof(op));
968 969 970 971 972 973 974 975 976 977
        op.cmd = XEN_V2_OP_SETVCPUMAP;
        op.domain = (domid_t) id;
        op.u.setvcpumap.vcpu = vcpu;
        op.u.setvcpumap.cpumap.bitmap = cpumap;
        op.u.setvcpumap.cpumap.nr_cpus = maplen * 8;
        ret = xenHypervisorDoV2Dom(handle, &op);
        if (munlock(cpumap, maplen) < 0) {
            virXenError(VIR_ERR_XEN_CALL, " release", maplen);
            ret = -1;
        }
978
    } else {
979 980 981
        cpumap_t xen_cpumap; /* limited to 64 CPUs in old hypervisors */
        uint64_t *pm = &xen_cpumap;
        int j;
982

983 984
        if ((maplen > (int)sizeof(cpumap_t)) || (sizeof(cpumap_t) & 7))
            return (-1);
985

986 987 988
        memset(pm, 0, sizeof(cpumap_t));
        for (j = 0; j < maplen; j++)
            *(pm + (j / 8)) |= cpumap[j] << (8 * (j & 7));
989 990

        if (hypervisor_version == 1) {
991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008
            xen_op_v1 op;

            memset(&op, 0, sizeof(op));
            op.cmd = XEN_V1_OP_SETVCPUMAP;
            op.u.setvcpumap.domain = (domid_t) id;
            op.u.setvcpumap.vcpu = vcpu;
            op.u.setvcpumap.cpumap = xen_cpumap;
            ret = xenHypervisorDoV1Op(handle, &op);
        } else if (hypervisor_version == 0) {
            xen_op_v0 op;

            memset(&op, 0, sizeof(op));
            op.cmd = XEN_V0_OP_SETVCPUMAP;
            op.u.setvcpumap.domain = (domid_t) id;
            op.u.setvcpumap.vcpu = vcpu;
            op.u.setvcpumap.cpumap = xen_cpumap;
            ret = xenHypervisorDoV0Op(handle, &op);
        }
1009 1010 1011
    }
    return(ret);
}
1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025
/**
 * virXen_getvcpusinfo:
 * @handle: the hypervisor handle
 * @id: the domain id
 * @vcpu: the vcpu to map
 * @cpumap: the bitmap for this vcpu
 * @maplen: the size of the bitmap in bytes
 *
 * Do a low level hypercall to change the pinning for vcpu
 *
 * Returns 0 or -1 in case of failure
 */
static int
virXen_getvcpusinfo(int handle, int id, unsigned int vcpu, virVcpuInfoPtr ipt,
1026
                    unsigned char *cpumap, int maplen)
1027 1028 1029 1030 1031 1032 1033
{
    int ret = -1;

    if (hypervisor_version > 1) {
        xen_op_v2_dom op;

        memset(&op, 0, sizeof(op));
1034 1035 1036 1037 1038 1039
        op.cmd = XEN_V2_OP_GETVCPUINFO;
        op.domain = (domid_t) id;
        op.u.setvcpumap.vcpu = vcpu;
        ret = xenHypervisorDoV2Dom(handle, &op);
        if (ret < 0)
            return(-1);
1040
        ipt->number = vcpu;
1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064
        if (op.u.getvcpuinfo.online) {
            if (op.u.getvcpuinfo.running) ipt->state = VIR_VCPU_RUNNING;
            if (op.u.getvcpuinfo.blocked) ipt->state = VIR_VCPU_BLOCKED;
        }
        else ipt->state = VIR_VCPU_OFFLINE;
        ipt->cpuTime = op.u.getvcpuinfo.cpu_time;
        ipt->cpu = op.u.getvcpuinfo.online ? (int)op.u.getvcpuinfo.cpu : -1;
        if ((cpumap != NULL) && (maplen > 0)) {
            if (mlock(cpumap, maplen) < 0) {
                virXenError(VIR_ERR_XEN_CALL, " locking", maplen);
                return (-1);
            }
            memset(&op, 0, sizeof(op));
            op.cmd = XEN_V2_OP_GETVCPUMAP;
            op.domain = (domid_t) id;
            op.u.setvcpumap.vcpu = vcpu;
            op.u.setvcpumap.cpumap.bitmap = cpumap;
            op.u.setvcpumap.cpumap.nr_cpus = maplen * 8;
            ret = xenHypervisorDoV2Dom(handle, &op);
            if (munlock(cpumap, maplen) < 0) {
                virXenError(VIR_ERR_XEN_CALL, " release", maplen);
                ret = -1;
            }
        }
1065
    } else {
1066 1067 1068 1069 1070
        int mapl = maplen;
        int cpu;

        if (maplen > (int)sizeof(cpumap_t))
            mapl = (int)sizeof(cpumap_t);
1071 1072

        if (hypervisor_version == 1) {
1073 1074 1075 1076 1077 1078 1079 1080 1081
            xen_op_v1 op;

            memset(&op, 0, sizeof(op));
            op.cmd = XEN_V1_OP_GETVCPUINFO;
            op.u.getvcpuinfo.domain = (domid_t) id;
            op.u.getvcpuinfo.vcpu = vcpu;
            ret = xenHypervisorDoV1Op(handle, &op);
            if (ret < 0)
                return(-1);
1082
            ipt->number = vcpu;
1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105
            if (op.u.getvcpuinfo.online) {
                if (op.u.getvcpuinfo.running) ipt->state = VIR_VCPU_RUNNING;
                if (op.u.getvcpuinfo.blocked) ipt->state = VIR_VCPU_BLOCKED;
            }
            else ipt->state = VIR_VCPU_OFFLINE;
            ipt->cpuTime = op.u.getvcpuinfo.cpu_time;
            ipt->cpu = op.u.getvcpuinfo.online ? (int)op.u.getvcpuinfo.cpu : -1;
            if ((cpumap != NULL) && (maplen > 0)) {
                for (cpu = 0; cpu < (mapl * 8); cpu++) {
                    if (op.u.getvcpuinfo.cpumap & ((uint64_t)1<<cpu))
                        VIR_USE_CPU(cpumap, cpu);
                }
            }
        } else if (hypervisor_version == 0) {
            xen_op_v1 op;

            memset(&op, 0, sizeof(op));
            op.cmd = XEN_V0_OP_GETVCPUINFO;
            op.u.getvcpuinfo.domain = (domid_t) id;
            op.u.getvcpuinfo.vcpu = vcpu;
            ret = xenHypervisorDoV0Op(handle, &op);
            if (ret < 0)
                return(-1);
1106
            ipt->number = vcpu;
1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120
            if (op.u.getvcpuinfo.online) {
                if (op.u.getvcpuinfo.running) ipt->state = VIR_VCPU_RUNNING;
                if (op.u.getvcpuinfo.blocked) ipt->state = VIR_VCPU_BLOCKED;
            }
            else ipt->state = VIR_VCPU_OFFLINE;
            ipt->cpuTime = op.u.getvcpuinfo.cpu_time;
            ipt->cpu = op.u.getvcpuinfo.online ? (int)op.u.getvcpuinfo.cpu : -1;
            if ((cpumap != NULL) && (maplen > 0)) {
                for (cpu = 0; cpu < (mapl * 8); cpu++) {
                    if (op.u.getvcpuinfo.cpumap & ((uint64_t)1<<cpu))
                        VIR_USE_CPU(cpumap, cpu);
                }
            }
        }
1121 1122 1123
    }
    return(ret);
}
1124
#endif /* !PROXY*/
1125

1126 1127 1128 1129 1130 1131 1132 1133 1134 1135
/**
 * xenHypervisorInit:
 *
 * Initialize the hypervisor layer. Try to detect the kind of interface
 * used i.e. pre or post changeset 10277
 */
int xenHypervisorInit(void)
{
    int fd, ret, cmd;
    hypercall_t hc;
1136
    v0_hypercall_t v0_hc;
1137
    xen_getdomaininfo info;
1138 1139

    if (initialized) {
1140
        if (hypervisor_version == -1)
1141 1142
            return(-1);
        return(0);
1143 1144
    }
    initialized = 1;
1145
    in_init = 1;
1146 1147 1148

    ret = open(XEN_HYPERVISOR_SOCKET, O_RDWR);
    if (ret < 0) {
1149
        hypervisor_version = -1;
1150 1151 1152 1153
        return (-1);
    }
    fd = ret;

1154 1155 1156 1157
    /*
     * The size of the hypervisor call block changed July 2006
     * this detect if we are using the new or old hypercall_t structure
     */
1158 1159 1160 1161 1162 1163 1164 1165
    hc.op = __HYPERVISOR_xen_version;
    hc.arg[0] = (unsigned long) XENVER_version;
    hc.arg[1] = 0;

    cmd = IOCTL_PRIVCMD_HYPERCALL;
    ret = ioctl(fd, cmd, (unsigned long) &hc);

    if ((ret != -1) && (ret != 0)) {
D
Daniel Veillard 已提交
1166
#ifdef DEBUG
1167
        fprintf(stderr, "Using new hypervisor call: %X\n", ret);
D
Daniel Veillard 已提交
1168
#endif
1169 1170 1171
        hv_version = ret;
        xen_ioctl_hypercall_cmd = cmd;
        goto detect_v2;
1172
    }
1173

1174 1175 1176 1177 1178 1179 1180 1181
    /*
     * check if the old hypercall are actually working
     */
    v0_hc.op = __HYPERVISOR_xen_version;
    v0_hc.arg[0] = (unsigned long) XENVER_version;
    v0_hc.arg[1] = 0;
    cmd = _IOC(_IOC_NONE, 'P', 0, sizeof(v0_hypercall_t));
    ret = ioctl(fd, cmd, (unsigned long) &v0_hc);
1182
    if ((ret != -1) && (ret != 0)) {
D
Daniel Veillard 已提交
1183
#ifdef DEBUG
1184
        fprintf(stderr, "Using old hypervisor call: %X\n", ret);
D
Daniel Veillard 已提交
1185
#endif
1186 1187
        hv_version = ret;
        xen_ioctl_hypercall_cmd = cmd;
1188
        hypervisor_version = 0;
1189
        goto done;
1190 1191
    }

1192 1193 1194 1195 1196
    /*
     * we faild to make any hypercall
     */

    hypervisor_version = -1;
1197 1198
    virXenError(VIR_ERR_XEN_CALL, " ioctl ", IOCTL_PRIVCMD_HYPERCALL);
    close(fd);
1199 1200 1201
    in_init = 0;
    return(-1);

1202
 detect_v2:
1203 1204 1205 1206 1207 1208 1209 1210 1211
    /*
     * The hypercalls were refactored into 3 different section in August 2006
     * Try to detect if we are running a version post 3.0.2 with the new ones
     * or the old ones
     */
    hypervisor_version = 2;
    /* TODO: one probably will need to autodetect thse subversions too */
    sys_interface_version = 2; /* XEN_SYSCTL_INTERFACE_VERSION */
    dom_interface_version = 3; /* XEN_DOMCTL_INTERFACE_VERSION */
1212
    if (virXen_getdomaininfo(fd, 0, &info) == 1) {
D
Daniel Veillard 已提交
1213
#ifdef DEBUG
1214
        fprintf(stderr, "Using hypervisor call v2, sys version 2\n");
D
Daniel Veillard 已提交
1215
#endif
1216
        goto done;
1217 1218 1219
    }
    hypervisor_version = 1;
    sys_interface_version = -1;
1220
    if (virXen_getdomaininfo(fd, 0, &info) == 1) {
D
Daniel Veillard 已提交
1221
#ifdef DEBUG
1222
        fprintf(stderr, "Using hypervisor call v1\n");
D
Daniel Veillard 已提交
1223
#endif
1224
        goto done;
1225 1226 1227 1228 1229 1230 1231 1232 1233 1234
    }

    /*
     * we faild to make the getdomaininfolist hypercall
     */

    hypervisor_version = -1;
    virXenError(VIR_ERR_XEN_CALL, " ioctl ", IOCTL_PRIVCMD_HYPERCALL);
    close(fd);
    in_init = 0;
1235 1236
    return(-1);

1237
 done:
1238
    close(fd);
1239
    in_init = 0;
1240 1241 1242
    return(0);
}

1243
#ifndef PROXY
1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255
/**
 * xenHypervisorRegister:
 *
 * Registers the xenHypervisor driver
 */
void xenHypervisorRegister(void)
{
    if (initialized == 0)
        xenHypervisorInit();

    virRegisterDriver(&xenHypervisorDriver);
}
1256
#endif /* !PROXY */
1257

1258 1259
/**
 * xenHypervisorOpen:
1260 1261 1262
 * @conn: pointer to the connection block
 * @name: URL for the target, NULL for local
 * @flags: combination of virDrvOpenFlag(s)
1263 1264 1265
 *
 * Connects to the Xen hypervisor.
 *
1266
 * Returns 0 or -1 in case of error.
1267
 */
1268
int
1269
xenHypervisorOpen(virConnectPtr conn, const char *name, int flags)
1270
{
1271 1272
    int ret;

1273 1274 1275
    if (initialized == 0)
        xenHypervisorInit();

1276
    if ((name != NULL) && (strcasecmp(name, "xen")))
1277 1278 1279 1280
        return(-1);

    conn->handle = -1;

1281
    ret = open(XEN_HYPERVISOR_SOCKET, O_RDWR);
1282
    if (ret < 0) {
1283
        if (!(flags & VIR_DRV_OPEN_QUIET))
1284 1285
            virXenError(VIR_ERR_NO_XEN, XEN_HYPERVISOR_SOCKET, 0);
        return (-1);
1286
    }
1287
    conn->handle = ret;
1288

1289
    return(0);
1290 1291 1292 1293
}

/**
 * xenHypervisorClose:
1294
 * @conn: pointer to the connection block
1295 1296 1297 1298 1299
 *
 * Close the connection to the Xen hypervisor.
 *
 * Returns 0 in case of success or -1 in case of error.
 */
1300
int
1301
xenHypervisorClose(virConnectPtr conn)
1302
{
1303 1304
    int ret;

1305
    if ((conn == NULL) || (conn->handle < 0))
1306
        return (-1);
1307

1308
    ret = close(conn->handle);
1309
    if (ret < 0)
1310 1311
        return (-1);
    return (0);
1312 1313 1314
}


1315
#ifndef PROXY
1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334
/**
 * xenHypervisorGetType:
 * @conn: pointer to the Xen Hypervisor block
 *
 * Get the version level of the Hypervisor running.
 *
 * Returns -1 in case of error, 0 otherwise. if the version can't be
 *    extracted by lack of capacities returns 0 and @hvVer is 0, otherwise
 *    @hvVer value is major * 1,000,000 + minor * 1,000 + release
 */
static const char *
xenHypervisorGetType(virConnectPtr conn)
{
    if (!VIR_IS_CONNECT(conn)) {
        virXenError(VIR_ERR_INVALID_CONN, __FUNCTION__, 0);
        return (NULL);
    }
    return("Xen");
}
1335
#endif
1336

1337 1338
/**
 * xenHypervisorGetVersion:
1339 1340
 * @conn: pointer to the connection block
 * @hvVer: where to store the version
1341 1342 1343
 *
 * Call the hypervisor to extracts his own internal API version
 *
1344
 * Returns 0 in case of success, -1 in case of error
1345
 */
1346 1347
int
xenHypervisorGetVersion(virConnectPtr conn, unsigned long *hvVer)
1348
{
1349 1350
    if ((conn == NULL) || (conn->handle < 0) || (hvVer == NULL))
        return (-1);
1351
    *hvVer = (hv_version >> 16) * 1000000 + (hv_version & 0xFFFF) * 1000;
1352
    return(0);
1353 1354
}

1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365
/**
 * xenHypervisorNumOfDomains:
 * @conn: pointer to the connection block
 *
 * Provides the number of active domains.
 *
 * Returns the number of domain found or -1 in case of error
 */
int
xenHypervisorNumOfDomains(virConnectPtr conn)
{
1366
    xen_getdomaininfolist dominfos;
1367 1368 1369 1370 1371 1372 1373
    int ret, nbids;
    static int last_maxids = 2;
    int maxids = last_maxids;

    if ((conn == NULL) || (conn->handle < 0))
        return (-1);

1374 1375
 retry:
    if (!(XEN_GETDOMAININFOLIST_ALLOC(dominfos, maxids))) {
1376
        virXenError(VIR_ERR_NO_MEMORY, _("allocating %d domain info"),
1377 1378
                    maxids);
        return(-1);
1379 1380
    }

1381 1382 1383
    XEN_GETDOMAININFOLIST_CLEAR(dominfos, maxids);

    ret = virXen_getdomaininfolist(conn->handle, 0, maxids, &dominfos);
1384

1385
    XEN_GETDOMAININFOLIST_FREE(dominfos);
1386 1387 1388 1389

    if (ret < 0)
        return (-1);

1390
    nbids = ret;
1391 1392 1393
    if (nbids == maxids) {
        last_maxids *= 2;
        maxids *= 2;
1394
        goto retry;
1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413
    }
    if ((nbids < 0) || (nbids > maxids))
        return(-1);
    return(nbids);
}

/**
 * xenHypervisorListDomains:
 * @conn: pointer to the connection block
 * @ids: array to collect the list of IDs of active domains
 * @maxids: size of @ids
 *
 * Collect the list of active domains, and store their ID in @maxids
 *
 * Returns the number of domain found or -1 in case of error
 */
int
xenHypervisorListDomains(virConnectPtr conn, int *ids, int maxids)
{
1414
    xen_getdomaininfolist dominfos;
1415 1416 1417 1418 1419 1420
    int ret, nbids, i;

    if ((conn == NULL) || (conn->handle < 0) ||
        (ids == NULL) || (maxids < 1))
        return (-1);

1421
    if (!(XEN_GETDOMAININFOLIST_ALLOC(dominfos, maxids))) {
1422
        virXenError(VIR_ERR_NO_MEMORY, "allocating %d domain info",
1423 1424
                    maxids);
        return(-1);
1425
    }
1426 1427

    XEN_GETDOMAININFOLIST_CLEAR(dominfos, maxids);
1428 1429
    memset(ids, 0, maxids * sizeof(int));

1430
    ret = virXen_getdomaininfolist(conn->handle, 0, maxids, &dominfos);
1431 1432

    if (ret < 0) {
1433
        XEN_GETDOMAININFOLIST_FREE(dominfos);
1434 1435 1436
        return (-1);
    }

1437
    nbids = ret;
1438
    if ((nbids < 0) || (nbids > maxids)) {
1439
        XEN_GETDOMAININFOLIST_FREE(dominfos);
1440 1441 1442 1443
        return(-1);
    }

    for (i = 0;i < nbids;i++) {
1444
        ids[i] = XEN_GETDOMAININFOLIST_DOMAIN(dominfos, i);
1445 1446
    }

1447
    XEN_GETDOMAININFOLIST_FREE(dominfos);
1448 1449 1450
    return (nbids);
}

1451
/**
1452 1453 1454
 * xenHypervisorGetDomMaxMemory:
 * @conn: connection data
 * @id: domain id
1455
 *
1456
 * Retrieve the maximum amount of physical memory allocated to a
1457
 * domain.
1458 1459 1460
 *
 * Returns the memory size in kilobytes or 0 in case of error.
 */
1461 1462
unsigned long
xenHypervisorGetDomMaxMemory(virConnectPtr conn, int id)
1463
{
1464
    xen_getdomaininfo dominfo;
1465 1466
    int ret;

1467
    if ((conn == NULL) || (conn->handle < 0))
1468 1469
        return (0);

1470 1471 1472 1473 1474 1475
    if (kb_per_pages == 0) {
        kb_per_pages = sysconf(_SC_PAGESIZE) / 1024;
	if (kb_per_pages <= 0) 
	    kb_per_pages = 4;
    }

1476
    XEN_GETDOMAININFO_CLEAR(dominfo);
1477

1478
    ret = virXen_getdomaininfo(conn->handle, id, &dominfo);
1479

1480
    if ((ret < 0) || (XEN_GETDOMAININFO_DOMAIN(dominfo) != id))
1481 1482
        return (0);

1483
    return((unsigned long) XEN_GETDOMAININFO_MAX_PAGES(dominfo) * kb_per_pages);
1484 1485
}

1486
#ifndef PROXY
1487 1488 1489
/**
 * xenHypervisorGetMaxMemory:
 * @domain: a domain object or NULL
1490
 *
1491 1492 1493 1494 1495 1496 1497 1498 1499 1500
 * Retrieve the maximum amount of physical memory allocated to a
 * domain. If domain is NULL, then this get the amount of memory reserved
 * to Domain0 i.e. the domain where the application runs.
 *
 * Returns the memory size in kilobytes or 0 in case of error.
 */
static unsigned long
xenHypervisorGetMaxMemory(virDomainPtr domain)
{
    if ((domain == NULL) || (domain->conn == NULL) ||
1501
        (domain->conn->handle < 0) || (domain->handle < 0))
1502 1503 1504 1505
        return (0);

    return(xenHypervisorGetDomMaxMemory(domain->conn, domain->handle));
}
1506
#endif
1507

1508
/**
1509 1510 1511
 * xenHypervisorGetDomInfo:
 * @conn: connection data
 * @id: the domain ID
1512
 * @info: the place where information should be stored
1513
 *
1514
 * Do an hypervisor call to get the related set of domain information.
1515 1516 1517 1518
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
1519
xenHypervisorGetDomInfo(virConnectPtr conn, int id, virDomainInfoPtr info)
1520
{
1521
    xen_getdomaininfo dominfo;
1522
    int ret;
1523
    uint32_t domain_flags, domain_state, domain_shutdown_cause;
1524 1525 1526 1527 1528 1529

    if (kb_per_pages == 0) {
        kb_per_pages = sysconf(_SC_PAGESIZE) / 1024;
	if (kb_per_pages <= 0) 
	    kb_per_pages = 4;
    }
1530

1531
    if ((conn == NULL) || (conn->handle < 0) || (info == NULL))
1532
        return (-1);
1533

1534
    memset(info, 0, sizeof(virDomainInfo));
1535
    XEN_GETDOMAININFO_CLEAR(dominfo);
1536

1537
    ret = virXen_getdomaininfo(conn->handle, id, &dominfo);
1538

1539
    if ((ret < 0) || (XEN_GETDOMAININFO_DOMAIN(dominfo) != id))
1540
        return (-1);
1541

1542 1543 1544
    domain_flags = XEN_GETDOMAININFO_FLAGS(dominfo);
    domain_state = domain_flags & 0xFF;
    switch (domain_state) {
1545 1546 1547 1548
	case DOMFLAGS_DYING:
	    info->state = VIR_DOMAIN_SHUTDOWN;
	    break;
	case DOMFLAGS_SHUTDOWN:
1549 1550 1551 1552 1553 1554 1555 1556 1557
            /* The domain is shutdown.  Determine the cause. */
            domain_shutdown_cause = domain_flags >> DOMFLAGS_SHUTDOWNSHIFT;
            switch (domain_shutdown_cause) {
                case SHUTDOWN_crash:
                    info->state = VIR_DOMAIN_CRASHED;
                    break;
                default:
                    info->state = VIR_DOMAIN_SHUTOFF;
            }
1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576
	    break;
	case DOMFLAGS_PAUSED:
	    info->state = VIR_DOMAIN_PAUSED;
	    break;
	case DOMFLAGS_BLOCKED:
	    info->state = VIR_DOMAIN_BLOCKED;
	    break;
	case DOMFLAGS_RUNNING:
	    info->state = VIR_DOMAIN_RUNNING;
	    break;
	default:
	    info->state = VIR_DOMAIN_NONE;
    }

    /*
     * the API brings back the cpu time in nanoseconds,
     * convert to microseconds, same thing convert to
     * kilobytes from page counts
     */
1577
    info->cpuTime = XEN_GETDOMAININFO_CPUTIME(dominfo);
1578 1579
    info->memory = XEN_GETDOMAININFO_TOT_PAGES(dominfo) * kb_per_pages;
    info->maxMem = XEN_GETDOMAININFO_MAX_PAGES(dominfo) * kb_per_pages;
1580
    info->nrVirtCpu = XEN_GETDOMAININFO_CPUCOUNT(dominfo);
1581
    return (0);
1582 1583
}

1584 1585 1586
/**
 * xenHypervisorGetDomainInfo:
 * @domain: pointer to the domain block
1587
 * @info: the place where information should be stored
1588
 *
1589
 * Do an hypervisor call to get the related set of domain information.
1590 1591 1592 1593 1594 1595 1596 1597
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
xenHypervisorGetDomainInfo(virDomainPtr domain, virDomainInfoPtr info)
{
    if ((domain == NULL) || (domain->conn == NULL) ||
        (domain->conn->handle < 0) || (info == NULL) ||
1598
        (domain->handle < 0))
1599
        return (-1);
1600

1601 1602 1603 1604
    return(xenHypervisorGetDomInfo(domain->conn, domain->handle, info));

}

1605
#ifndef PROXY
1606 1607
/**
 * xenHypervisorPauseDomain:
1608
 * @domain: pointer to the domain block
1609 1610 1611 1612 1613 1614
 *
 * Do an hypervisor call to pause the given domain
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
1615
xenHypervisorPauseDomain(virDomainPtr domain)
1616
{
1617 1618
    int ret;

1619
    if ((domain == NULL) || (domain->conn == NULL) ||
1620
        (domain->conn->handle < 0) || (domain->handle < 0))
1621 1622
        return (-1);

1623
    ret = virXen_pausedomain(domain->conn->handle, domain->handle);
1624
    if (ret < 0)
1625 1626
        return (-1);
    return (0);
1627 1628 1629 1630
}

/**
 * xenHypervisorResumeDomain:
1631
 * @domain: pointer to the domain block
1632 1633 1634 1635 1636 1637
 *
 * Do an hypervisor call to resume the given domain
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
1638
xenHypervisorResumeDomain(virDomainPtr domain)
1639
{
1640 1641
    int ret;

1642
    if ((domain == NULL) || (domain->conn == NULL) ||
1643
        (domain->conn->handle < 0) || (domain->handle < 0))
1644 1645
        return (-1);

1646
    ret = virXen_unpausedomain(domain->conn->handle, domain->handle);
1647
    if (ret < 0)
1648 1649
        return (-1);
    return (0);
1650 1651 1652 1653
}

/**
 * xenHypervisorDestroyDomain:
1654
 * @domain: pointer to the domain block
1655 1656 1657 1658 1659 1660
 *
 * Do an hypervisor call to destroy the given domain
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
1661
xenHypervisorDestroyDomain(virDomainPtr domain)
1662
{
1663 1664
    int ret;

1665
    if ((domain == NULL) || (domain->conn == NULL) ||
1666
        (domain->conn->handle < 0) || (domain->handle < 0))
1667 1668
        return (-1);

1669
    ret = virXen_destroydomain(domain->conn->handle, domain->handle);
1670
    if (ret < 0)
1671 1672
        return (-1);
    return (0);
1673 1674
}

1675 1676
/**
 * xenHypervisorSetMaxMemory:
1677
 * @domain: pointer to the domain block
1678 1679 1680 1681 1682 1683 1684
 * @memory: the max memory size in kilobytes.
 *
 * Do an hypervisor call to change the maximum amount of memory used
 *
 * Returns 0 in case of success, -1 in case of error.
 */
int
1685
xenHypervisorSetMaxMemory(virDomainPtr domain, unsigned long memory)
1686
{
1687 1688
    int ret;

1689
    if ((domain == NULL) || (domain->conn == NULL) ||
1690
        (domain->conn->handle < 0) || (domain->handle < 0))
1691 1692
        return (-1);

1693
    ret = virXen_setmaxmem(domain->conn->handle, domain->handle, memory);
1694
    if (ret < 0)
1695 1696
        return (-1);
    return (0);
1697
}
1698
#endif /* PROXY */
1699

1700
#ifndef PROXY
1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713
/**
 * xenHypervisorSetVcpus:
 * @domain: pointer to domain object
 * @nvcpus: the new number of virtual CPUs for this domain
 *
 * Dynamically change the number of virtual CPUs used by the domain.
 *
 * Returns 0 in case of success, -1 in case of failure.
 */

int
xenHypervisorSetVcpus(virDomainPtr domain, unsigned int nvcpus)
{
1714
    int ret;
1715

1716
    if ((domain == NULL) || (domain->conn == NULL) ||
1717 1718
        (domain->conn->handle < 0) || (domain->handle < 0) ||
        (nvcpus < 1))
1719
        return (-1);
1720 1721 1722

    ret = virXen_setmaxvcpus(domain->conn->handle, domain->handle, nvcpus);
    if (ret < 0)
1723
        return (-1);
1724
    return (0);
1725 1726 1727 1728 1729 1730 1731 1732
}

/**
 * xenHypervisorPinVcpu:
 * @domain: pointer to domain object
 * @vcpu: virtual CPU number
 * @cpumap: pointer to a bit map of real CPUs (in 8-bit bytes)
 * @maplen: length of cpumap in bytes
1733
 *
1734 1735 1736 1737 1738 1739 1740 1741 1742
 * Dynamically change the real CPUs which can be allocated to a virtual CPU.
 *
 * Returns 0 in case of success, -1 in case of failure.
 */

int
xenHypervisorPinVcpu(virDomainPtr domain, unsigned int vcpu,
                     unsigned char *cpumap, int maplen)
{
1743
    int ret;
1744

1745
    if ((domain == NULL) || (domain->conn == NULL) ||
1746 1747
        (domain->conn->handle < 0) || (domain->handle < 0) ||
        (cpumap == NULL) || (maplen < 1))
1748 1749
        return (-1);

1750 1751 1752
    ret = virXen_setvcpumap(domain->conn->handle, domain->handle, vcpu,
                            cpumap, maplen);
    if (ret < 0)
1753
        return (-1);
1754
    return (0);
1755
}
1756
#endif
1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770

/**
 * virDomainGetVcpus:
 * @domain: pointer to domain object, or NULL for Domain0
 * @info: pointer to an array of virVcpuInfo structures (OUT)
 * @maxinfo: number of structures in info array
 * @cpumaps: pointer to an bit map of real CPUs for all vcpus of this domain (in 8-bit bytes) (OUT)
 *	If cpumaps is NULL, then no cupmap information is returned by the API.
 *	It's assumed there is <maxinfo> cpumap in cpumaps array.
 *	The memory allocated to cpumaps must be (maxinfo * maplen) bytes
 *	(ie: calloc(maxinfo, maplen)).
 *	One cpumap inside cpumaps has the format described in virDomainPinVcpu() API.
 * @maplen: number of bytes in one cpumap, from 1 up to size of CPU map in
 *	underlying virtualization system (Xen...).
1771
 *
1772 1773 1774 1775 1776
 * Extract information about virtual CPUs of domain, store it in info array
 * and also in cpumaps if this pointer is'nt NULL.
 *
 * Returns the number of info filled in case of success, -1 in case of failure.
 */
1777
#ifndef PROXY
1778 1779
int
xenHypervisorGetVcpus(virDomainPtr domain, virVcpuInfoPtr info, int maxinfo,
1780
                      unsigned char *cpumaps, int maplen)
1781
{
1782
    xen_getdomaininfo dominfo;
1783 1784
    int ret;

1785
    virVcpuInfoPtr ipt;
1786
    int nbinfo, i;
1787

1788 1789 1790 1791
    if ((domain == NULL) || (domain->conn == NULL) ||
        (domain->conn->handle < 0) || (domain->handle < 0) ||
        (info == NULL) || (maxinfo < 1) ||
        (sizeof(cpumap_t) & 7))
1792
        return (-1);
1793
    if ((cpumaps != NULL) && (maplen < 1))
1794
        return -1;
1795 1796

    /* first get the number of virtual CPUs in this domain */
1797 1798 1799
    XEN_GETDOMAININFO_CLEAR(dominfo);
    ret = virXen_getdomaininfo(domain->conn->handle, domain->handle,
                               &dominfo);
1800

1801
    if ((ret < 0) || (XEN_GETDOMAININFO_DOMAIN(dominfo) != domain->handle))
1802
        return (-1);
1803
    nbinfo = XEN_GETDOMAININFO_CPUCOUNT(dominfo) + 1;
1804 1805 1806
    if (nbinfo > maxinfo) nbinfo = maxinfo;

    if (cpumaps != NULL)
1807
        memset(cpumaps, 0, maxinfo * maplen);
1808

1809 1810
    for (i = 0, ipt = info; i < nbinfo; i++, ipt++) {
        if ((cpumaps != NULL) && (i < maxinfo)) {
1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822
            ret = virXen_getvcpusinfo(domain->conn->handle, domain->handle, i,
                                      ipt,
                                      (unsigned char *)VIR_GET_CPUMAP(cpumaps, maplen, i),
                                      maplen);
            if (ret < 0)
                return(-1);
        } else {
            ret = virXen_getvcpusinfo(domain->conn->handle, domain->handle, i,
                                      ipt, NULL, 0);
            if (ret < 0)
                return(-1);
        }
1823 1824 1825
    }
    return nbinfo;
}
1826 1827 1828 1829 1830 1831 1832 1833 1834 1835
#endif

/*
 * Local variables:
 *  indent-tabs-mode: nil
 *  c-indent-level: 4
 *  c-basic-offset: 4
 *  tab-width: 4
 * End:
 */